822
Bioconjugate Chem. 2000, 11, 822−826
Anchor-Chain Molecular System for Orientation Control in Enzyme Immobilization Wen-Hai Shao,† Xian-En Zhang,* Hong Liu, and Zhi-Ping Zhang Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, 430071, P. R. China
Anthony E. G. Cass Department of Biochemistry, Imperial College of Science, Technology & Medicine, South Kensington, London SW7 2AY, U.K. Received March 20, 2000; Revised Manuscript Received July 13, 2000
An anchor-chain molecular system was constructed for controlled orientation and high activity in enzyme immobilization. A streptavidin recognition peptide (streptag) coding sequence was fused to the 3′ end of the phoA gene, which codes for E. coli alkaline phosphatase (EAP). Both the wild-type (WT) and the Asp-101 f Ser (D1O1S) mutant were modified with the streptag sequence with or without the insertion of a flexible linker peptide [-(Gly-Ser)5-] coding sequence. The fused genes were cloned into the vector pASK75 and expressed in the periplasm of the host cell Escherichia coli SM547. The proteins were released by osmotic shock and purified by ion-exchange chromatography. Enzyme activities of all proteins were measured spectrophotometrically with F-nitrophenyl phosphate as the substrate. Specific activities of D101S-streptag and D101S-linker-streptag enzymes were increased 25- or 34-fold over the WT, respectively. These fusion proteins were then immobilized on microtiter plates through streptag-streptavidin binding reaction. After immobilization, the D101S-linker-streptag enzyme displayed the highest residual activity and the ratio of enzyme activities of the linker to nonlinker enzymes was 8.4. These results show that the addition of a linker peptide provides a spacer so as to minimize steric hindrance between the enzyme and streptavidin. The method provides a solution for controlled enzyme immobilization with high recover activity, which is especially important in construction of biosensors, biochips, or other biodevices.
INTRODUCTION
Understanding the properties of immobilized proteins is critical to the optimal design of biosensors, bioseparations, bioreactors, and the production of bioelectronic devices will undoubtedly demand much better control of the positioning and assembly of biomolecular materials as well as high retention of biological activity. Protein patterning, the defined spatial localization of molecules on a surface, is a powerful approach to generate molecular arrays for analytical or bioelectronic applications and similarly demands a high degree of molecular control over the orientation of the molecules on the surface. In most circumstances, however, immobilization chemistries require a significant amount of enzyme. Likewise, immobilized enzymes very often show reduced reaction rates and product yields (1). This is mainly due to the multisite attachment, multiple orientations, and steric hindrance that result from nonspecific chemical crosslinking of protein to surface. Attempts have been made to solve the problems by more precisely defined proteinsurface interations. Examples include site-specific attachment to gold surfaces by introducing a single cysteine residue into the cysteine-free enzyme (2), orientational control of enzymes by the use of nonporous hydroxyalkyl methacrylate solid support (3), and oriented immobilization of antibodies by the use of immobilized protein A or G (4). * To whom correspondence should be addressed. Phone/Fax: +86 (0)27 87641492. E-mail:
[email protected]. † Current address: Department of Biology, Xuzhou Normal University, Jiangsu province, 221009, P. R. China.
One of the most promising methods is to use fusion protein technology. Fusion proteins, comprising separate protein sequences have been widely used as aids to expression and purification of recombinant proteins. The gene for the protein of interest is fused to a sequence that codes for a protein or peptide, which in turn binds to a specific molecular structure on the surface. Many of these expression/purification systems are commercially available, and their basic characteristics have been reviewed (5). These fusion sequences can also be used to immobilize proteins on surfaces for sensing or molecular electronic applications. The efficiency of immobilization depends on the carrier, immobilization method, and initial enzyme concentration. Gritsch et al. (6) described the binding of a poly (His) fusion protein to surfaces coated with metalchelating self-assembled lipid monolayers. Recently, a method was described for photopatterning surfaces with biotin (7) and biotinylated EAP bound via a streptavidin bridge. An alternative to biotinylation is to use a streptavidin binding peptide sequence (streptag). The streptag sequence was created by selection from a genetic fusion peptide library by its ability to bind to streptavidin in a specific and reversible manner (8). Streptavidin is usually preferred to avidin, due to its lower nonspecific binding characteristics and high resistance to denaturation (9). An EAP-streptag fusion protein system has been characterized (10) and, although the activity of the fusion protein was lower than WT, the work demonstrated the potential use of the avidin-biotin system for protein immobilization under mild conditions. Escherichia coli alkaline phosphatase (EC 3.1.3.1, EAP) is a homodimeric zinc metalloprotein that has been
10.1021/bc000029s CCC: $19.00 © 2000 American Chemical Society Published on Web 11/01/2000
Anchor-Chain Molecular System
Bioconjugate Chem., Vol. 11, No. 6, 2000 823
Figure 1. Strategy of linker insertion. An asterisk (*) indicates modified amino acid residues, which leads to elimination of the restriction enzyme site at the 5′ termination of linker. The insertion fragment must be designed to be subject to the enzyme PstI. The synthesized oligos contain an overlap at both ends to adapt to the catalytic reaction of PstI.
extensively characterized (11-13). Its broad substrate specificity means that the enzyme activity can be readily measured when immobilized using either colorimetric or fluorescent end-point detection. The major drawback of the enzyme in analytical applications is its relatively low activity. Several point mutations in the vicinity of the active site have been shown to alter the enzymatic activity. In particular, data from the introduction of a serine hydroxyl group at position 101 suggests that the mutant enzyme achieves a quicker catalytic turnover by allowing faster product release (14, 15). Since direct fusion of streptag to EAP led to a 50% loss of enzyme activity (7), we sought to improve on this by inserting a linker peptide between the C-terminus of EAP and the streptag sequence. This was thought to provide a sufficiently flexible spacer for both proteins to remain in native conformation. This may improve the control of the orientation of enzyme during immobilization, so resulting in a high recovery activity. EXPERIMENTAL PROCEDURE
Materials. Agar, agarose, ampicillin (Amp), F-nitrophenyl phosphate (pNPP), and streptavidin were purchased from Sigma Chemical Co. Tryptone and yeast extract were obtained from England OXOID. Restriction endonucleases were products of either GIBCO BRL or Promega Co. Kits used for the extraction of plasmid and the isolation of DNA fragments from agarose gels were purchased from QIAGEN. T4 DNA ligase was obtained from GIBCO BRL. A stirred cell system used for protein concentration was purchased from FILTRON. A Titertek Uniscan microplate reader was used to measure the activity of the immobilized enzymes. Purification was accomplished using a Q Sepharose (Sigma) ion-exchange chromatography column. The single strand oligonucleotide coding for the linker peptide was commercially obtained from
Sangon (Shanghai, China). The E. coli strain SM547 and the plasmid pEK48 kindly donated by professor Kantrowitz and pASK75 was purchased from Biometra. Construction of Fusion Proteins. The EAP-streptag fusion proteins were generated by cloning phoA into the plasmid pASK75 (10). The linker peptide oligonucleotide was annealed into double-strand DNA prior to mixing with pASK75, which had been previously cut with PstI (Figure 1). The mixture was treated with T4 DNA ligase overnight at 14 °C, followed by transforming into competent E. coli SM547 cells and plating on LB medium plates containing 100 µg/mL Amp and 40 µg/mL 5-bromo4-chloro-3-indolyl phosphate (BCIP) (w/v). Twelve of the blue colonies (displaying EAP activity) were selected, and plasmid DNA was isolated. Each of the plasmid candidates was first checked for the correct size by agarose gel electrophoresis and then analyzed by PCR. A plasmid, which contained the correct inserted linker peptide sequence, was isolated. The construction of plasmids containing the mutation was accomplished by cutting both plasmid pASK75 and the D1O1S mutant plasmid pEK48 (X.-E. Zhang et al., unpublished results) with the restriction enzymes Tth111 I and BstxI. Digestion with these enzymes produced a fragment of the phoA gene containing the mutational site. The strategy was to replace the fragment from the mutant plasmid pEK48. To ensure that the fragment from the WT gene was not able to religate, the pASK75 backbone was removed from an agarose gel after electrophoresis using QIAquick Gel Extraction Kit. Transformation into competent E. coli SM547 cells was performed. The mixture was plated on LB medium plate. The isolation of the target plasmid was accomplished as before. DNA sequencing results also confirmed that the linker peptide and the new mutational sites were introduced. Protein Purification. Cells were grown at 37 °C in LB medium containing 100 µg/mL Amp with shaking
824 Bioconjugate Chem., Vol. 11, No. 6, 2000
until the absorbance at 600 nm reached 0.5. Expression was then induced by the addition of isoprophylthio-βD-galactoside (IPTG) (for cells containing pEK48) or Tetracycline (for cells containing pASK75) to a final concentration of 1 mM or 0.2 µg/mL (w/v), respectively. After the bacterial cells were grown for 20 h, they were harvested by centrifugation. The fusion protein was released by osmotic shock according to the method of Brockman and Heppel (16). The supernatant containing the fusion proteins were combined and purified using Q Sepharose ion-exchange chromatography. Initially, the column was washed with 20% ethanol (v/v) and then equilibrated with 10 mM Tris (TMZP buffer, 100 mM Tris containing 1 × 10-3 M MgCl2, 1 × 10-4 M NaH2PO4, 3.1 × 10-3 M NaN3, and 1 × 10-5 M ZnSO4, pH 7.4). The protein was applied to the column in the same buffer and unbound protein was eluted with equilibration buffer. The protein was eluted using a salt gradient in the TMZP buffer, and 8 mL fractions were collected and tested for EAP activity. Proteins were concentrated using the FILTRON stirred cell system. Final protein preparations were stored at -20 °C in TMZP buffer. Protein concentration and purity were determined by using Bio-Rad Assay Kit with bovine serum albumin as the standard and SDS-PAGE, respectively. Kinetic Assays. EAP is catalytically active at relatively high pH values. The enzyme is most active at pH 10.0, and accordingly, all reactions were carried out at that pH in 1 M Tris buffer. FNPP was used as an EAP substrate to measure the enzymatic activities spectrophotometrically and, hence, determine the kinetic parameters. The release of F-nitrophenolate (FNP) was monitored at 405 nm. A total of 3 µL of enzyme solution was added to the microtiter wells containing 33 µL of 3 M Tris-HCl (pH 10.0) and 64 µL of FNPP (50 mM TrisHCI buffer; 1 mM MgCl2; pH 10.0). The reaction mixture was incubated for 1 min at 25 °C. The absorbance at 405 nm in each well of the plate was then read after subtraction of the blank value for each dilution. The activity assays after immobilization were performed as described below. A total of 100 µL of 100 µg/ mL streptavidin was applied to each well of the microtiter plate and incubated for 1 h at 4 °C. The plate was washed three times with distilled water and subsequently blocked for 4 h in 3% BSA solution (w/v). After washing, 100 µL of 10 µg/mL EAP fusion protein was added and incubated for 30 min. Unbound enzyme was removed by washing three times. Finally, 100 µL of FNPP solution (20 mM FNPP and 1 M Tris-HCl, pH 10.0) was added. The color signal was developed for 2 min. The absorption values were determined using a microplate reader and data were reported as A405 values for the sample wells after subtraction of the blank value. The measurements were repeated three times. Protein Immobilization. One microliter of 100 µg/ mL streptavidin was applied to each well of the glass slide and incubated for 4 h at 4 °C. After removal of the solution, the plate was blocked with 3% BSA solution (w/ v), and incubated for 4 h. After washing, 1 µL of 10 µg/ mL EAP fusion protein was applied and incubated for 2 h. Unbound enzymes were removed by washing three times with distilled water. Finally, 1 µL of NBT/BCIP (another EAP substrate, diluted 4:4:25 in buffer supplied by the HuaMei manufacturer) was added. Color change was seen within 10 min. Unless stated otherwise, the experiments were performed at room temperature (25 °C).
Shao et al.
Figure 2. Elution pattern of the proteins. Peak 1, EAPstreptag; peak 2, EAP; peak 3, EAP-linker- streptag. RESULTS
Alteration of Elution Peak. All proteins had the same elution peak profile, but required different ionic strengths of the eluting buffer. The WT and fusion protein containing streptag can be eluted from the column by using 0 to 0.1 M NaCI concentration gradient solution. However, the WT was eluted at 0.058 M NaCI, while the WT-streptag fusion protein was eluted at 0.046 M NaCl. The insertion of linker peptide required a higher ionic strength and was eluted at 0.17 M NaCl (Figure 2). All samples were eluted at the same flow rate and dissolved in the same buffer. Kinetic Properties of EAP Fusion Proteins. Table 1 lists the EAP fusion proteins that were constructed and the kinetic constants. Replacement of Asp 101 by Ser in EAP has a large effect on kcat. The kcat and Km of WT EAP streptag fusion protein are similar to the wild-type. The fusion protein D101S-streptag has about a 9- and 11-fold increase of kcat value over the WT and WT-streptag fusion protein, respectively. The insertion of the linker into the D101S-streptag fusion protein resulted in slight increase of kcat value. The kcat of fusion protein D101S-linkerstreptag is 15 times higher than the fusion protein WTstreptag. The kcat/Km ratio for each of the enzymes is almost identical with the corresponding value for the WT enzyme, since the increment in kcat is compensated for by increase of 23- and 40-fold in the Km of the D101Sstreptag and D101S-linker-streptag fusion enzymes, respectively (Table 1). Therefore, despite the higher catalytic activity of the mutant enzymes each is still as effective a catalyst as the WT enzyme. The D101S-streptag shows about 40-fold Km increases over the WT. The introduction of a linker led to a 0.6fold decreased Km value relative to the nonlinker mutant, whereas it has the highest specific activity value. This result suggests that the D101S-linker-streptag fusion protein has the highest catalytic efficiency and a higher affinity to the substrate comparing with the fusion protein D101S-streptag. The activity assay after immobilization showed that D101S-linker-streptag also developed the highest residual activity. After immobilization (in microtiter plate), the ratio of residue enzyme activities of the D101S-linker-streptag and D101Sstreptag was 8.4, but the ratio of the enzyme activities of the same proteins before immobilization is 2.0. Figure 3 showed the enzyme activity of the fusion proteins after immobilization and nonimmobilization with or without the insertion of the linker peptide. The catalytic activities of the fusion proteins in solution were higher than that in immobilized conditions. But enzymes without linker peptide gave a sharply decreased catalytic activity, while
Anchor-Chain Molecular System
Bioconjugate Chem., Vol. 11, No. 6, 2000 825
Table 1. Kinetic Parameters of E. coli EAP and Fusion Proteins enzyme
specific activity (U,a pH 10.0, 1 M Tris)
Km (µM, pH 10.0, 1 M Tris)
kcatb (s-1)
kcat/Km (S-1 µM-1)
WT WT-streptag D101S-streptag WT-linker-streptag D101S-linker-streptag
53 47 1422 87 1799
89 86 3509 94 2080
27.5 21.0 243.0 13.6 308.0
0.30 0.24 0.07 0.14 0.15
a U ) µmol of the substrate FNPP hydrolyzed/mg of protein/min. b The k cat values are calculated per active site from the Vmax using a dimer molecular weight of 94 000 (11). The buffer system is 1.0 M Tris-HCI, pH 10.0. The MW of streptag and linker peptide is about 1100, respectively.
Figure 3. Effect of linker on the degree of residual activity on the polystyrene surface. The fusion proteins were assayed in the absence and presence of linker. Enzyme activity was expressed relative to the activity of the fusion protein in absence of linker. Data represent the mean of three duplicates from one experiment. Key 1, immobilized D101S-linker-streptag (0) and D101S-streptag (9); key 2, nonimmobilized D101S-linkerstreptag (0) and D101S-streptag (9); key 3, immobilized WTlinker-streptag (0) and WT-streptag (9); key 4, nonimmobilized WT-linker-streptag (0) and WT-streptag (9).
the enzymes with the insertion of linker peptide had a higher recover activity. Immobilization on Glass Support. Immobilization of the fusion proteins was performed on the surface of a glass slide with four duplicates of each sample. The slide with immobilized proteins was thoroughly washed with distilled water followed by adding the NBT/BCIP and incubated for 10 min. Color then developed. Relative recovery activities of the fusion proteins were evaluated according to the depth of the color spots. As shown in Figure 4, the sequence of depth of the color was D101Slinker-streptag, D101S-streptag, and WT-linker-streptag. Results from the variance analysis and t-test also showed that the color among three candidates has extremely notable discrimination, the average margins between arbitrary two of the three candidates (25.75; 35.25; 9.5) are greater than the value of t0.01 × SD (8.52). No detectable signal was observed when WT, which was used as an experiment control, was spotted on the glass surface by the same procedure. DISCUSSION
The active sites of enzyme are less rigid and more sensitive to slight changes of conformation during immobilization. General immobilization process, e.g., chemical cross-linking or covalent binding, usually results in alterations in the level of enzyme activity. There are numerous chemical groups on the surface of the enzyme for immobilizing enzyme onto solid supports. For example, there are 40 lysine residues in the EAP structure, most of them are located on the surface of the enzyme with exposure of their amido groups, providing random chemical binding sites. Some of the amido groups are right in the active sites, which may result in poor
Figure 4. Immobilization of the proteins on glass chip. The positions of the three positive reactions (each loaded four duplicates) and three negative reactions are indicated. Experimental detail is in the text. (A) Denotes the detection of EAP; (B) denotes the detection of EAP-linker-streptag; (C) denotes the detection without substrate; (D) denotes the detection of D101S-streptag; (E) denotes the detection of D101S-linkerstreptag; (F) denotes the detection without the streptavidin.
recovery activity if chemical reagents are used for immobilization. The conception of using EAP-streptag system is to avoid random immobilization as that the C-terminal of the enzyme protein domain, where the streptag fused to, is the unique binding site, leaving the active sites opening toward the bulk solution. The streptag here acts as an anchor. The problem of less activity of the EAP-streptag is overcome by insertion of the linker peptide. This linker peptide is composed of glycine and serine. Glycine is the simplest amino acid, with no ionizable carbon side chain. The absence of a side chain gives the polypeptide backbone at glycine residues much greater conformational flexibility. The side chain of serine is short and hydrophilic due to the presence of a hydroxyl group. The peptide composed of these two amino acids forms a stable R-helix (as viewed by Swiss Protein Data Bank Viewer, SPDBV) in water solution (Figure 5). When the immobilized EAP was secluded away from the support surface by using this long and hydrophilic anchor-chain or spacer arm, it exhibits minimal steric hindrances that could be promoted by the streptag and the proximity of the support surface. The data showed the evidence that catalytic activities of the fusion proteins with linker are much higher than that of the fusion proteins without linker. Charges on the streptag and the relative position of the peptide and the EAP in aqueous solution affect the elution during ion-exchange chromatography. The streptag peptide contains 10 amino acids (SAWRHPQPGG), of which, only two (R and H) are basic amino acids. The side chain of arginine consists of three hydrophobic
826 Bioconjugate Chem., Vol. 11, No. 6, 2000
Shao et al.
streptag. The proposed anchor-chain method could serve as a model to build other fusion protein systems and is optimistic to achieve better orientation control of enzyme immobilization, which is potentially useful in high quality construction of biosensors, biochips, and biodevices. ACKNOWLEDGMENT
This work was supported by Chinese National Nature Science Foundation (no. 39770221) and Joint Project between the Royal Society and the Chinese Academy of Sciences (no. Q174). We thank Dr. Catherine Halliwell for helpful discussions. LITERATURE CITED
Figure 5. Three-dimensional structure of the EAP (A), EAPstreptag (B) and EAP-linker-streptag (C). The linker peptide shows R-helix at the C-terminal of EAP.
methylene groups and the strong basic δ-guanido group, of which pK′R is 12.48 and is ionized over almost entire pH range, i.e., it is positively charged in our experimental condition (pH 7.4). Histidine is one of the strongest bases with a pK′R of 6.20. It is not surprising that EAP-streptag fusion protein needs a lower ionic strength to be eluted from column as the tag has two extra positively amino acids (R and H) than WT. But what is surprising is the large different elution ionic strength between EAPstreptag and EAP-linker-streptag. It is not due to any charge in the linker as the linker is composed of amino acids (G and S) that are neutral at pH 7.4. It is thought that this large difference is due to the linker altering the position of the tag relative to the EAP, the co-effect of the linker, and streptag peptide on the surface of the enzyme, or some unknown interaction between proteins that may lead to charges exposing or covering. That needs to be further verified. CONCLUSION
We have successfully obtained a higher recover activity by the insertion of a linker peptide between EAP and
(1) Tischer, W., and Kasche, V. (1999) Immobilized Enzymes: Crystal or Arriers? Trends Biotechnol. 17 (8), 326-335. (2) Huang, W., Wang, J., Bhattacharyya, D., and Bachas, L. G. (1997) Improving the Activity of Immobilized Subtilisin by Site-Specific Attachment to Surfaces. Anal. Chem. 69 (22), 4601-4607. (3) Bilkova, Z., Mazurova, J., Churacek, J., Horak, D., and Turkova, J. (1999) OrientedImmobilization of Chymotrypsin by Use of Suitable Antibodies Coupled to a Nonporous Solid Support. J. Chromatogr A. 852 (1), 141-149. (4) Turkova, J. (1999) Oriented Immobilization of Biologically Active Proteins as aTool for Revealing Protein Interactions and Function. J. Chromatogr. B 722 (1-2), 11-31. (5) Nygren, P. A., Stahl, S., and Uhln, M. (1994) Engineering Proteins to FacilitateBioprocessing. Trends Biotechnol. 12, 184-188. (6) Gritsch, S., Neumaier, K., Schmitt, L., and Tampd, R. (1995) Engineered FusionMolecules at Chelator Lipid Interfaces Imaged by Reflection Interference Contrast Microscopy (RICM). Biosens. Bioelectron. 10, 805-812. (7) Hengsakul, M., and Cass, A. E. G. (1996) Protein Patterning with aPhotoactivatable Derivative of Biotin. Bioconjugate Chem. 7, 249-254. (8) Schmidt, T. G. M., Koepke, J., Frank, R., and Skerra, A. (1996) MolecularInteraction between the strep-tag Affinity Peptide and its Congate Target Streptavidin. J. Mol. Biol. 255, 753-766. (9) Savage, M. D. (1996) An Introduction to Avidin-Biotin Technology and Options for Biotinylation. BioMethods, Vol. 7, p 2, Germany. (10) Hengsakul, M., and Cass, A. E. G. (1997) Alkaline Phosphatase-Strep Tag Fusion Protein Binding to Streptavidin: Resonant Mirror Studies. J. Mol. Biol. 266, 621-632. (11) Broadshaw, R. A., Cancedda, F., Ericsson, L. H., Newman, P. A., Piccoli, S. P., Schesinger, K., and Walsh, K. A. (1981) Amino Acid Sequence of Escherichia coli Alkaline Phosphatase. Proc. Natl. Acad. Sci. U.S.A. 78, 3473-3477. (12) Chang, N. C., Wun-Jing, K., and Ellson, Y. C. (1986) Nucleotide Sequence of theAlkaline Phosphatase Gene of Escherichia coli. Gene 44, 121-125. (13) Kim, E. E., and Wyckoff, H. W. (1991) Reaction Mechanism of AlkalinePhosphatase Based on Crystal Structure, TwoMetal Ion Catalysis. J. Mol. Biol. 218, 449-464. (14) Liqing, C., Neidhart, D., Kohlbrenner, W. M., Mandecki, W., Bell, S., Sowadski, J., and Abad-Zapatero, C. (1992) 3-D Structure of a Mutant (AsplOlfSer) of E. coli Alkaline Phoshatase with Higher Catalytic Activity. Protein Eng. 5 (7), 605-610. (15) Brennan, C. A., Christlanson, K., Fleur, M. A. L., and Mandecki, W. (1995) A Molecular Sensor System Based on Genetically Engineered Alkaline Phosphatase. Biochemistry 92, 5783-5787. (16) Brockman, R. W., and Heppel, L. A. (1968) On the Localization of AlkalinePhosphatase and Cyclic Phosphodiesterase in Escherichia coli. Biochemistry 7, 2554-2562.
BC000029S
